Averaging values within bins

Hi,

I have the two arrays x and y (attached), which I plot with plot(x,y). This is creating "bins" of width 5:

What is the fastest way to average the values within each such bin, such that we end up with one value/bin, especially if we don't want to pre-define bin width?

Thanks

[Stephen23]

"This is creating "bins" of width 5:"

Are you referring to the x-tick marks and x-tick labels? Note that the data has not been processed according to those tick marks. They are simply an artifact of the axes.

If you want to bin data, then you will need to do this yourself, e.g. using a histogram function, or accumarray, or something similar.

[MiauMiau]

I don't want to have a histogram in the classical sense: I want the mean value within a bin (not the count). Would the histogram function help nevertheless?

[Stephen23]

"Would the histogram function help nevertheless?"

Yes, because they can also return the indices.

Answer 1

[Bruno Luong]

load('y.mat')
load('x.mat')

plot(x,y)
edges = (0:5:50);
[~,~,loc]=histcounts(x,edges);
meany = accumarray(loc(:),y(:))./accumarray(loc(:),1);
xmid = 0.5*(edges(1:end-1)+edges(2:end));

figure
plot(x,y)
hold on; 
plot(xmid,meany,'r')

[AnnieLieseMari]

Hi there!

I'm working with a very similar data set and need some help with my code. I know this was posted and answered a while ago, but I would really appreciate the help.

After using the same code you provided, but with my data set and edges of (4.25:0.5:17.75), I am getting an error saying "First input SUBS must contain positive integer subscripts."

I have a data set of x and y values that is upwards of 2000+ data points. It's super big and when I zip it, the Help Center is still saying it's too big. Please let me know if you'd like me to provide a sample of maybe 20 or so data points to help out.

I think the main reason this is happening is that my data has NaN values in it. Though, I'm not sure how to remove those. Of course, if I remove the x values, I have to preserve the y values so the other pairs still match up.

[Bruno Luong]

Simply remove the bad data:

keep = isfinite(x) & isfinit(y);
x = x(keep);
y = y(keep);
... do the rest

Make sure the extrem edges cover entirely your data.

Answer 2

[Steven Lord]

Do you want to use histogram or histcounts to bin the data, or are you looking for something more of a moving average, where the value of the averaged signal at a particular time is the mean of the data for that time and some data points before and after that time? If the latter, see movmean.